ovn-kubernetes CNI request failed with status 400

[Gjonni]

Describe the bug

I have found several times that the OVN kubernetes service no longer releases ip to new pods, as a solution I tried to delete both the ovnkube-master- * and ovnkube-node * pods but the problem persisted.
So I decided to enter the master nodes and delete the files / var / lib / onv / etc / * (ovn db) then I restarted the ovnkube- * pods again and I solved the problem.
Unfortunately, periodically I find myself in the same situation.
on the various pod logs I find the presence of the CNI request failed with status 400 error I think the problem is very similar or the same as https://bugzilla.redhat.com/show_bug.cgi?id=1994810.
Is there a definitive FIX?

Version
IPI installation - 4.9.0-0.okd-2022-02-12-140851

How reproducible
on my environment it shows up about once a month. I honestly don't know how to reproduce it.
my environment makes use of several machinesets with additional networks

Log bundle

[Gjonni]

@vrutkovs I'm not quite sure it's a discussion because the CNI request failed with status 400 error I find it on different installations but it does not always generate the problem of the OVN database, but that's okay as a discussion :-).

[vrutkovs]

Moving this to discussion until log bundle is provided, so that we could confirm its indeed an OKD specific bug

[Gjonni]

perfect thanks